Low-Rank Variational Quantum Algorithm for the Dynamics of Open Quantum Systems

Sara Santos,Xinyu Song,Vincenzo Savona

2024-03-09

Abstract:The simulation of many-body open quantum systems is key to solving numerous outstanding problems in physics, chemistry, material science, and in the development of quantum technologies. Near-term quantum computers may bring considerable advantage for the efficient simulation of their static and dynamical properties, thanks to hybrid quantum-classical variational algorithms to approximate the dynamics of the density matrix describing the quantum state in terms of an ensemble average. Here, a variational quantum algorithm is developed to simulate the real-time evolution of the density matrix governed by the Lindblad master equation, under the assumption that the quantum state has a bounded entropy along the dynamics, entailing a low-rank representation of its density matrix. The algorithm encodes each pure state of the statistical mixture as a parametrized quantum circuit, and the associated probabilities as additional variational parameters stored classically, thereby requiring a significantly lower number of qubits than algorithms where the full density matrix is encoded in the quantum memory. Two variational Ansätze are proposed, and their effectiveness is assessed in the simulation of the dynamics of a 2D dissipative transverse field Ising model. The results underscore the algorithm's efficiency in simulating the dynamics of open quantum systems in the low-rank regime with limited quantum resources on a near-term quantum device.

Quantum Physics

What problem does this paper attempt to address?

The paper attempts to address the problem of efficiently simulating the dynamics of open quantum systems on near-term quantum devices (NISQ devices). Specifically, the authors propose a novel variational quantum algorithm to directly solve the Lindblad master equation for simulating the real-time evolution of open quantum systems. This algorithm is based on low-rank approximation, representing mixed states as statistical mixtures of pure states, and encoding each pure state through parameterized quantum circuits, while the probabilities are stored as additional variational parameters in a classical computer. This approach significantly reduces the number of required qubits, thereby lowering computational complexity.

Additionally, the paper proposes two different variational forms (Ansatz) and validates their effectiveness in the dynamical simulation of a 2D dissipative transverse field Ising model. The results indicate that the algorithm can effectively simulate the dynamics of open quantum systems under low-rank conditions with limited quantum resources. This method holds significant implications for the development of physics, chemistry, materials science, and quantum technology.